LAUSR

Standard sparse algorithms always assume that all the azimuths lie in prior discretized spatial grids in direction-of-arrival (DOA) estimations. However, this assumption may lead to poor performance in practice owing to the spatial arbitrariness of true azimuths. Several techniques have been proposed to overcome this off-the-grid issue, but the performance of these techniques is not satisfactory (i.e., they are either inaccurate or computationally expensive). In this letter, we propose a post-processing algorithm, called dynamic parameterized $\ell _{1}$ -regulation, which efficiently provides compressed-sensing-based single-snapshot DOA estimations. The advantages of our proposed algorithm are verified from our numerical results.